LED ERP vs. current

While searching for optical RX-TX systems that would carry voice, I stumbled across an article of the only Italian HAM (AFAIK) that has documented optical experiments, I4VIL. He mentions that LED efficiency at some point decreases even if the current flow increases (in Italian).

I managed to replicate his observations using my MCW TX and the simple OPT301 RX, just the detector without the active bandpass filter. I monitored the received signal strength by feeding the detector output to my laptop MIC input. Spectran software did the measurements. The TX was aimed at the ceiling, as well as the OPT301, without any lens.

I monitored the total circuit input current, not the one flowing into the LED. The MCW TX draws about 6 mA @12V in stand-by, so you can figure out how much current really flows into the LED. But at 100mA average when transmitting, 6mA are negligible.


The Y-axis shows dB reading on Spectran for the 488 Hz line, relative to the first dot (17 mA); the X-axis is the total current flowing into the circuit.

It is obvious that the emitted light reaches the maximum power somewhere between 80 and 100 mA. Going further just heats the diode.

Delta ERP is +17 dB. That would be almost 3 S-points in RF terms.

Interesting, isn't it?